An improved dead reckoning method for mobile robot with redundant odometry information

Abstract

Dead reckoning is an important method for mobile robot. If its accuracy can be improved, navigation tasks will be simplified. The mobile robot concerned in this paper is with five wheels, in which there are dual driving wheels and a castor. A measurement wheel with encoder is fixed beside each driving wheel. There are two encoders fixed on the castor. One measures the castor's movement, and another indicates the angle of yawing. A new kind of kinematics equations derived from the robot's turning radius and angle of movement trajectory is presented. We can get a group of turning radius and angle from the data of four encoders, in which there are four different values for radius and six different values for angle. The estimated values of radius and angle can be gotten by fusion using fuzzy algorithm. The improved dead reckoning method has the advantages of simplicity, less computation cost, cheapness and good performance in uneven road. The simulation shows its effectiveness.